My machine programming company's website: http://merly.ai
My Machine Programming & Technology YouTube Channel (subscribe & stay updated)
Keynote at LADSIOS (co-located with VLDB '21): "Machine Programming and the Future of Software Development"
New demo of one of our production quality MP systems: AutoPerf
Our team, joint w/ MIT & Microsoft, won two awards at SIGMOD '21!
Keynote @ Penn's PRECISE 2019 Industry Day: "Machine Programming: The Future of Autonomy"
NeurIPS '22, ICLR'22, PLDI'22, CGO'21, NeurIPS'21, AIDB'21, PACT'21, FSE'21, OOPSLA'21, MAPS'21 (SC chair), ICML'21, USENIX ATC'21, ICLR'21, MLSys'21, NeurIPS'20, MAPL'20 (SC chair), JPDC'20, aiDM '20, TheWebConf'20, MLSys'20, PACT'19 (SRC), SysML'19, MAPL'18 (general chair)
Brief Biographical Sketch
In December 2021, I founded Merly, Inc., a Silicon Valley start-up aiming at disrupting software development. I am currently Merly's Chief Executive Officer (CEO) and Chief Scientist. Merly aims to (i) improve the rate at which we develop software while concurrently (ii) improving its quality. We achieve this by employing a variety of automation techniques -- otherwise known as machine programming -- such as deep neural networks and formal methods. Learn more here (https://merly.ai)!
Previously, I founded and led the Machine Programming Research group at Intel Labs. Machine programming (MP) is a new field of research that uses automation to improve the rate at which we develop software (e.g., the time it takes a developer to write, maintain, and test code) and improve its associated quality characteristics (e.g., performance, correctness, security, maintainability, etc.). We generally consider MP as a fusion of machine learning and formal methods, which rely heavily on programming languages and systems. We provide a brief overview of MP in our “Three Pillars of Machine Programming” vision paper (see Armando Solar-Lezama's website for a deeper dive).
In academia, I am an instructor at Stanford University, where I teach the "Introduction to Machine Programming" graduate computer science course.
I have 40+ peer reviewed papers, 60+ issued patents, and 110+ patents pending. I've been lucky enough to have been invited to give talks at places like Berkeley, BMW, DARPA, IBM Research, MIT, Penn, Stanford, UCLA, University of Washington, VMWare, and Wharton, amongst others. I've had the tremendous honor to give keynote addresses at places like VLDB (LADSIOS), University of Pennsylvania, the US Department of Energy, and MIT. My team's research has been highlighted by venues like The Wall Street Journal, DeepLearning.ai, Communications of the ACM, MIT Technology Review, The New York Times, and many others.
53rd patent issued: "Methods and apparatus for detecting a side channel attack using hardware performance counters" (11,188,643)
52nd patent issued: "Methods, systems, and articles of manufacture to perform heterogeneous data structure selection via programmer annotations" (11,188,324)
51st patent issued: "Methods, systems, articles of manufacture and apparatus for code review assistance for dynamically typed languages" (11,157,384)
Keynote address at LADSIOS (co-located with VLDB '21): "Machine Programming and the Future of Software Development"
[Milestone] 50th patent issued: "Methods and apparatus to detect side-channel attacks"
49th patent issued: "Methods and apparatus to automatically generate code for graphical user interfaces"
48th patent issued:: "Efficient sharing and compression expansion of data across processing systems"
Our team, Machine Programming Research (MPR), won two awards at SIGMOD '21!
Accepted to the KDD Workshop on Programming Language Processing: "A Survey on Semantic Parsing for Machine Programming"
47th patent issued: "Methods and apparatus to improve utilization of a heterogeneous system executing software" (11,036,477)
46th patent issued: "Methods and apparatus to validate data communicated by a vehicle" (11,024,180)
45th patent issued: "Methods and apparatus for recommending computer program updates utilizing a trained model" (11,003,444)
Accepted to the 2021 ACM SIGPLAN Machine Programming Symposium (MAPS): "ControlFlag: A Self-Supervised Idiosyncratic Pattern Detection System for Software Control Structures"
Accepted to the 2021 ACM SIGPLAN Machine Programming Symposium (MAPS): "Predictive Locality Optimization for Higher-Order Tensor Computations"
Accepted to 2021 GECCO Workshop on Evolutionary Computation Software Systems (EvoSoft): "AI Programmer: Autonomously Creating Software Programs Using Genetic Algorithms"
44th patent issued: "Extend GPU/CPU coherency to multi-GPU cores" (10,956,330)
43rd patent issued: "Methods and apparatus to detect memory leaks in computing systems" (10,956,298)
42nd patent issued: "Neural network optimization mechanism" (10,929,749)
41st patent issued: "Methods and apparatus for runtime multi-scheduling of software executing on a heterogeneous system" (10,908,884)
Upcoming keynote address @ MIT's DSAIL 2021 virtual retreat: "A Glimpse Into Machine Programming @ Intel Labs"
40th patent issued: "Compute optimization for deep neural networks" (10,902,547)
39th patent issued: "Coordination and increased utilization of graphics processors during inference" (10,891,707)
A video on our ControlFlag system
38th patent issued: "Systems and methods for determining a configuration of a microarchitecture" (10,853,554)
MS advisor (University of Pennsylvania): Brad MacDonald -> Tesla
MS co-advisor (University of Pennsylvania): Celine Lee -> Intel Labs, then PhD student @ Cornell
PhD committee member (Lehigh University): PanteA Zardoshti -> Microsoft Research
PhD committee member (University of Washington): Maaz Ahmad -> Adobe Research